Engineering high affinity humanized anti-p185HER2/anti-CD3 bispecific F(ab')2 for efficient lysis of p185HER2 overexpressing tumor cells.

We previously constructed a humanized anti-p185HER2/anti-CD3 bispecific antibody variant, BsF(ab')2 v1 which retargets the cytotoxic activity of human T cells in vitro against human breast tumor cells which overexpress the p185HER2 product of the HER2/neu (c-erbB-2) protooncogene. Subsequently we identified an improved anti-CD3 variant, v9, which binds to T cells with approx. 100-fold higher affinity than the original variant, v1. Here we demonstrate that BsF(ab')2 v9 is more potent than BsF(ab')2 v1 in stimulating the proliferation of both resting peripheral blood lymphocytes (PBL) and IL-2-activated, long-term cultured T lymphocytes (ATL). In addition, at low concentrations (0.01-1 ng/ml) BsF(ab')2 v9 is much more efficient than BsF(ab')2 v1 in directing lysis of p185HER2-overexpressing tumor cells by IL-2 activated PBL. In contrast, at higher concentration BsF(ab')2 v9 and BsF(ab')2 v1 have similar potency in retargeted cytotoxicity. At BsF(ab')2 v9 concentrations of > or = 1 ng/ml the susceptibility of p185HER2-expressing tumor cells to lysis is apparently independent of the level of p185HER2 expression. At lower concentrations of BsF(ab')2 v9 and/or lower ratios of effector to target cells the extent of lysis is reduced, in some cases improving the selectivity of lysis of high p185HER2 expressors over low expressors. Thus selection of a high affinity anti-CD3 arm is likely important in the design of BsF(ab')2 for retargeting the cytotoxicity of T cells to tumors. The dose of BsF(ab')2 v9 in any future clinical evaluation will require optimization to maximize anti-tumor efficacy whilst minimizing potential toxicity against normal tissue expressing p185HER2.